Intelligent Multi-Agent Collaboration Model for Smart Home IoT Security

While the Internet of Things (IoT) continue to extend deeper into the daily lives of people, the domain of the smart home offers a unique need for security. Traditional firewalls and antivirus are not sufficient to protect the connected home from security threats, and to date, there have been limited solutions provided for this problem. This paper explores the recent works in this area and presents a new approach towards securing smart home networks through multi-agent collaboration. The model uses Beliefs, Desires, and Intentions (BDI) architecture for intelligent agent decision making, as well as a multi-agent collaboration model for achieving mutual security goals within the smart home network. For initial proof of concept, we provide a use case demonstrating the coordination of the threat response decision between operational availability and security risk agents as a qualitative coalitional game. This model can also be extended to other areas such as threat modeling, vulnerability scanning, and patching, as well as more advanced threat engagement and distraction. Throughout, we also consider operational goals of convenience and availability as required for a usability perspective within the smart home environment.